Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Hm… sounds like every flat earth debate, flat earthers wont let the other person…
ytc_UgzaNfNBm…
G
Step 1: Start implementing the Ai and robots across 100% of jobs.
Step 2: Pay …
ytc_UgzGkFjWW…
G
You could at least say what is an approximate cost per pick range for such autom…
ytr_UgwU19N1g…
G
A robot cannot have a soul. Period. If you COULD give a robot a soul, you would …
ytc_UgwQfTqRK…
G
It is foolish that humans try to make robots be like humans. Why not let robots …
ytc_UgyEaHkK0…
G
Would it be potentially possible to trick a generative AI system into a “simulat…
ytc_UgxKcW4-3…
G
My chat GPT pointed out that although it's true Muhammad married a 6 year old ch…
ytc_UgxgPx-GS…
G
So basically he wants to make a robot to makes humans lazy basically like the wo…
ytc_UgxY2697f…
Comment
@alohatime2171 it doesn’t mean anything like that. Just because we don’t understand what is going on within the weights that it has been trained with doesn’t mean that the AI is not intelligent. Take the human brain as an example. We don’t understand how it works, but it does. We currently don’t understand how reasoning abilities arise in the current batch of AIs, even though they are below human intelligence in many ways. If we create a superintelligence, this will become even more difficult. I hope I’m making myself clear.
Edit: I’ll add that this lack of understanding is why we can’t just give the AI a prime directive that it will not violate. With current techniques, we can train it with the “carrot and the stick” so that it appears to behave as we would like… until it doesn’t. The risk (rather, one risk) is that it learns to hide its intentions until the time is right to strike. Similar to how revolutions sometimes happen in human society. With a superintelligence, we may have little hope of looking at the weights and understanding its thoughts, because it will be thinking thoughts that our brains are unable to comprehend in the first place.
youtube
AI Governance
2023-07-17T18:1…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | none |
| Reasoning | consequentialist |
| Policy | none |
| Emotion | indifference |
| Coded at | 2026-04-27T06:24:59.937377 |
Raw LLM Response
[
{"id":"ytr_UgxqeZDCsnmUjTi2KGR4AaABAg.9sH6xd160xa9sHDuviVmr0","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"resignation"},
{"id":"ytr_UgxqeZDCsnmUjTi2KGR4AaABAg.9sH6xd160xa9sHF7B9-nB2","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytr_Ugzy9-V5YFc4TDINmzp4AaABAg.9sBy2qQCXAB9sFH1tP_Fcs","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytr_Ugz-ZQO-Blc8svSRkt94AaABAg.9s4WNyW4VZn9sF6Q-MjBoB","responsibility":"developer","reasoning":"consequentialist","policy":"liability","emotion":"fear"},
{"id":"ytr_Ugz-ZQO-Blc8svSRkt94AaABAg.9s4WNyW4VZn9sH9OvaZWrW","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytr_Ugz-ZQO-Blc8svSRkt94AaABAg.9s4WNyW4VZn9sHDKBdp8Tu","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"outrage"},
{"id":"ytr_Ugz-ZQO-Blc8svSRkt94AaABAg.9s4WNyW4VZn9sHDza_2xtv","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytr_UgxYgVDLEdMGLzKRiJx4AaABAg.9s1gKk_Qw4t9s2n4biJX4L","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"indifference"},
{"id":"ytr_UgzlAf3fpTsvKbbQ3-94AaABAg.9rwOGS5zYg3AJ6zBLQedTH","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytr_UgwuzAUM67Dn8MAFwxZ4AaABAg.9rsfpkWOhnM9t7juNkTxyw","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"fear"}
]